Commit Graph

14 Commits

Author SHA1 Message Date
Craig Topper
e4fd907e72 Use uint16_t to store register overlaps to reduce static data.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152001 91177308-0d34-0410-b5e6-96231b3b80d8
2012-03-04 10:43:23 +00:00
Craig Topper
b6632ba380 Use uint16_t instead of unsigned to store registers in reg classes. Reduces static data size.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151998 91177308-0d34-0410-b5e6-96231b3b80d8
2012-03-04 10:16:38 +00:00
Craig Topper
015f228861 Use uint16_t to store registers in callee saved register tables to reduce size of static data.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151996 91177308-0d34-0410-b5e6-96231b3b80d8
2012-03-04 03:33:22 +00:00
Jakob Stoklund Olesen
f79b489571 Add missing static
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151396 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-24 21:52:44 +00:00
Jakob Stoklund Olesen
27bc818eaf Add a -stress-regalloc=<N> option.
This will limit all register classes to N registers in order to stress
test register allocation.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151379 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-24 18:34:20 +00:00
Jakob Stoklund Olesen
f39031b360 Detect proper register sub-classes.
Some instructions require restricted register classes, but most of the
time that doesn't affect register allocation. For example, some
instructions don't work with the stack pointer, but that is a reserved
register anyway.

Sometimes it matters, GR32_ABCD only has 4 allocatable registers. For
such a proper sub-class, the register allocator should try to enable
register class inflation since that makes more registers available for
allocation.

Make sure only legal super-classes are considered. For example, tGPR is
not a proper sub-class in Thumb mode, but in ARM mode it is.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@136981 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-05 21:28:14 +00:00
Jakob Stoklund Olesen
79c890f64f Add TargetRegisterInfo::getRawAllocationOrder().
This virtual function will replace allocation_order_begin/end as the one
to override when implementing custom allocation orders. It is simpler to
have one function return an ArrayRef than having two virtual functions
computing different ends of the same array.

Use getRawAllocationOrder() in place of allocation_order_begin() where
it makes sense, but leave some clients that look like they really want
the filtered allocation orders from RegisterClassInfo.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133170 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-16 17:42:25 +00:00
Jakob Stoklund Olesen
687397c013 Include callee-saved registers in debug output.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132899 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-13 03:26:42 +00:00
Jakob Stoklund Olesen
7c48913af6 Don't try to be clever, just preserve the target's allocation order.
The order of registers returned by getCalleeSavedRegs is used to lay out
the fixed stack slots for CSRs. Some targets like their CSRs used from
one end, and some targets want them used from the other end.

When computing an allocation order, simply preserve the relative
ordering of CSRs that the target specifies in its allocation order.
Reordering CSRs would break some targets, ARM in particular.

We still place volatiles before the CSRs, providing slightly better
results with different calling conventions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132680 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-06 16:36:30 +00:00
Jakob Stoklund Olesen
d365fa9415 Preserve the original ordering when a CSR has multiple aliases.
Previously, these aliases would be ordered alphabetically. (BH, BL)

Print out the computed allocation orders.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132580 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-03 20:34:50 +00:00
Jakob Stoklund Olesen
6edf90b8a7 Just use a SmallVector.
I was confused whether new uint8_t[] would zero-initialize the returned
array, and it seems that so is gcc-4.0.

This should fix the test failures on darwin 9.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132500 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-02 22:22:43 +00:00
Benjamin Kramer
19e1f633af Start with a zeroed CSRNum map.
Found by valgrind.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132457 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-02 12:07:44 +00:00
Jakob Stoklund Olesen
ab5ceacbc1 Initialize members to fix problem found by valgrind.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132456 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-02 05:43:49 +00:00
Jakob Stoklund Olesen
491a13691d Add a RegisterClassInfo class that lazily caches information about
register classes.

It provides information for each register class that cannot be
determined statically, like:

- The number of allocatable registers in a class after filtering out the
  reserved and invalid registers.

- The preferred allocation order with registers that overlap callee-saved
  registers last.

- The last callee-saved register that overlaps a given physical register.

This information usually doesn't change between functions, so it is
reused for compiling multiple functions when possible.  The many
possible combinations of reserved and callee saves registers makes it
unfeasible to compute this information statically in TableGen.

Use RegisterClassInfo to count available registers in various heuristics
in SimpleRegisterCoalescing, making the pass run 4% faster.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132450 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-02 02:19:35 +00:00